```html Building a Merch Store Between Rides

Building a Merch Store Between Rides: How AI Collapsed the Distance Between Idea and Checkout

Another day, another build. This one happened in parking lots and side streets across Vancouver, squeezed into the gaps between Uber rides.

PlayBunny.ai is a print-on-demand hoodie store I built in about half a dozen sessions using nothing but my phone and an AI coding assistant. The honest version of that story is more interesting than the romanticized one, so here's what actually happened.

The Printify Debugging Chronicles

The whole thing started with Printify, a print-on-demand service that's essentially the spiritual successor to mail-order catalogs—except everything runs on REST APIs. I fired off requests to their endpoints looking for available products and sizing data. The AI returned beautiful-looking code right up to the point where it invented an endpoint that doesn't exist. Printify's real endpoint has a specific nesting of shipping and line items that took three debugging sessions to get right.

This is the part that doesn't make it into the tweet threads. AI-assisted development is genuinely powerful, but it has no sense of what's real. It generates plausible-looking boilerplate with complete confidence. Your job is to know the difference.

The Mockup Problem

Printify's API returns product mockups that looked like they were rendered in 2003—low resolution, flat lighting, nothing you'd put on a product page. So I used AI image generation to produce proper mockups instead. The results were good enough to sell on.

The Stack

Stripe handles payments. Printify handles fulfillment. A small Python server on my VPS on Hostinger coordinates between them via webhooks. When someone buys a hoodie, Stripe fires a webhook, my server receives it, and a Printify order gets placed automatically. I'm not touching anything.

I'm also experimenting with Phantom wallet integration for Solana payments—the connect button is live on the site, with full checkout coming soon. The appeal is obvious: no card data, instant settlement, no chargebacks.

The Beekeeper's Lesson

In 1852, a Philadelphia minister named Lorenzo Langstroth discovered that if you leave exactly 9.5mm between honeycombs, bees won't seal the frames with wax. Just enough room to move, not enough to build. He called it bee space. Every modern beehive is built around that one measurement.

I think about that when I'm debugging. There's always one small constraint that unlocks the whole system. For Printify it was the endpoint structure. For Stripe it was the webhook signature. Find the bee space and everything else follows.

What It Actually Means

Print-on-demand margins are thin—around 15-20% after fees, which means volume matters. But that's not really the point yet. The point is that a working e-commerce operation with global fulfillment now takes half a dozen sessions instead of months. The barriers have never been lower.

PlayBunny isn't going to make me rich. But it works, it ships, and I built it between rides.

```

codebuddy.tech